Oracle Cloud Infrastructure Documentation

Resetting DR Configuration After a Failover

Issue: A failover plan execution does not clean up resources in the primary DR protection group or configure DR member properties to prepare for a DR transition in the reverse direction.

Solution: After a failover plan execution has succeeded, you must manually perform this cleanup and reverse the configuration of properties to prepare the DR configuration for switchover or failover in the reverse direction when needed.
Note

Performing the manual cleanup step listed here will require you to refresh all the existing DR plans. After performing a refresh, you must verify all the DR plans.

Perform the following cleanup and reconfiguration tasks manually at the end of a failover plan execution:

Before you begin cleanup for each individual member type described here, ensure that the role and the state of the old primary (new standby) DR protection group is correct. The role of the new standby DR protection group should be Standby, and its lifecycle state should be Active. You can edit the DR protection group’s role by clicking Edit next to the Role field in the DR protection group’s information pane. You can reset the DR protection group’s lifecycle state by selecting Reset State in the More Actions menu.

Volume Group Members

For each volume group member in the old primary DR protection group, perform the following steps:

  1. Disable replication for each original volume group member in the old primary DR protection group.
  2. Remove all the original volume group members from the old primary DR protection group.
  3. Terminate all the volume groups that were removed from the old primary DR protection group in the previous step.
  4. Configure the volume group replication in the reverse direction (from current primary region to current standby region) for each new volume group member added to the current primary DR protection group.
  5. If required, configure any additional member properties for each new volume group member in the current primary DR protection group, such as Destination backup policy, Destination Customer Managed Keys, and so on.

File System members

For each file system member in the old primary DR protection group, perform the following steps:

  1. Disable replication for each original file system member in the old primary DR protection group.
  2. Remove all the original file system members from the old primary DR protection group.
  3. Terminate all the file systems that were removed from the old primary DR protection group in the previous step.
  4. Configure the file system replication in the reverse direction (from current primary region to current standby region) for each new file system member added to the current primary DR protection group.
  5. If required, configure any additional member properties for each new file system member in the current primary DR protection group, such as Destination snapshot policy, Destination Customer Managed Keys, and so on.

Object Storage members (buckets)

For each object storage bucket member in the old primary DR protection group, perform the following steps:

  1. Disable replication for each original object storage bucket in the old primary DR protection group.
  2. Remove all the original object storage buckets from the old primary DR protection group.
  3. Do not terminate these removed object storage buckets.
  4. Configure the object storage bucket replication in the reverse direction (from current primary region to current standby region) for each object storage bucket that exists in the current primary DR protection group.

Moving Compute Instance members

For each moving compute instance member in the old primary DR protection group, perform the following steps:

  1. Detach and terminate all the block volumes attached to moving instance members in the old primary DR protection group.
  2. Remove all the original moving compute instance members from the old primary DR protection group.
  3. Terminate all the moving compute instances that were removed from the old primary DR protection group in the previous step. Ensure that you also delete the attached boot volumes when terminating these compute instances.

Non-moving Compute Instance members

For each non-moving compute instance member in the old primary DR protection group, perform the following steps:

  1. In the old primary DR protection group, detach and terminate all the block volumes that were configured for attach-detach operations. You can find the list of these block volumes in the Block volumes tab of the Advanced options section.
  2. In the old primary DR protection group , unmount all the file systems that were configured for mount-unmount file system operations. You can find the list of these file systems in the File systems tab of the Advanced options section.
  3. In the old primary DR protection group, remove the existing Mount target details and reconfigure the Mount target properties in the File system tab in the Advanced options section.
  4. In the old primary DR protection group, remove the existing attach-detach properties for the block volumes in the Block volumes tab in the Advanced options section, and reconfigure them using block volumes information from the new primary DR protection group.

Database members (Base Database, Exadata Database [all types], Autonomous Container Database)

For each database member in the old primary DR protection group, perform the following steps:

  1. Verify using the OCI Database console that the new primary database has a Primary role, and the old primary database has a Disabled Standby role.
  2. Perform a Data Guard reinstate operation on each database members in the old primary DR protection group. See Perform Database Switchover and Failover and Reinstate a Database for additional details.

Load Balancer and Network Load Balancer members

For each load balancer and network load balancer member in the old primary DR protection group, perform the following steps:

  1. For backend sets pointing to moving instances, delete all the backends from the backend set.
  2. For backend sets pointing to non-moving instances, mark the backend set as offline.

OKE Cluster members

For each OKE Cluster member in the old primary DR protection group, perform the following steps:

  1. Login to the OKE Cluster and remove the namespace named brie. This will disable the backup of the OKE cluster.
  2. Login to the OKE Cluster and manually delete all the namespaces that were configured for DR. Ensure that all the resources associated with these namespaces are also deleted. Examples of these associated resources are: pods, deployments, persistent volumes, CRDs, and so on.

Parent topic: Troubleshooting